biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[MoE Refactor] Separate Router into OO Classes (#30623)

Browse Source 
Signed-off-by: Bill Nell <bnell@redhat.com>
This commit is contained in:
bnellnm
2026-01-18 11:40:49 -05:00
committed by GitHub
parent 2f03035a61
commit 327a02d8db
45 changed files with 1750 additions and 684 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
tests/kernels/moe/test_flashinfer_moe.py
View File

@@ -12,6 +12,7 @@ from tests.kernels.quantization.nvfp4_utils import (
from tests.kernels.utils import torch_moe
from vllm import _custom_ops as ops
from vllm.config import ParallelConfig, VllmConfig, set_current_vllm_config
from vllm.model_executor.layers.fused_moe import fused_topk
from vllm.model_executor.layers.fused_moe.flashinfer_cutlass_moe import (
FlashInferExperts,
is_valid_flashinfer_cutlass_fused_moe,
@@ -19,7 +20,6 @@ from vllm.model_executor.layers.fused_moe.flashinfer_cutlass_moe import (
from vllm.model_executor.layers.fused_moe.flashinfer_cutlass_prepare_finalize import (
create_flashinfer_prepare_finalize,
)
from vllm.model_executor.layers.fused_moe.fused_moe import fused_topk
from vllm.model_executor.layers.fused_moe.modular_kernel import FusedMoEModularKernel
from vllm.platforms import current_platform
from vllm.utils.flashinfer import has_flashinfer_cutlass_fused_moe